Heart failure with preserved ejection fraction: A clustering approach to a heterogenous syndrome

Heart failure with preserved ejection fraction (HFpEF) is a complex syndrome at the crossroads of multiple co-morbidities; there is no valid treatment for this condition. Defining new phenotypes could play a role in improving treatment and prognosis.

To identify groups with different pathophysiologies by applying a clustering approach to a multicentric cohort of patients with HFpEF.

A total of 538 patients from the multicentre KaRen study were included. Accurate clinical, biological and ultrasound data are available, with a mean follow-up of 28 months. Based on a clustering analysis, the population was separated into groups based on 55 variables, comparing distribution of deaths and hospitalizations between groups.

Three clusters were identified from 356 analysable patients (mean age 76.1±9.31 years; 43.5% men): cluster 1 (n=128) comprised overweight, relatively young men at high cardiovascular risk, in sinus rhythm, with altered renal function; cluster 2 (n=134) comprised women, most of whom had conserved left ventricular function; cluster 3 (n=94) had the highest incidence of mitral regurgitation, atrial remodelling and rhythm disorders. There were no significant differences, only a trend towards early mortality in cluster 3.

Clustering analysis seems to be effective at individualizing subgroups with different physiopathologies in HFpEF. The clinical relevance of these phenotypes needs to be studied, and may concern treatment strategy more than prognostic differences.